_ Git - youtube-dl/blob - youtube_dl/extractor/generic.py

   1 # encoding: utf-8
   2
   3 import os
   4 import re
   5
   6 from .common import InfoExtractor
   7 from ..utils import (
   8     compat_urllib_error,
   9     compat_urllib_parse,
  10     compat_urllib_request,
  11     compat_urlparse,
  12
  13     ExtractorError,
  14     smuggle_url,
  15     unescapeHTML,
  16     unified_strdate,
  17     url_basename,
  18 )
  19 from .brightcove import BrightcoveIE
  20
  21
  22 class GenericIE(InfoExtractor):
  23     IE_DESC = u'Generic downloader that works on some sites'
  24     _VALID_URL = r'.*'
  25     IE_NAME = u'generic'
  26     _TESTS = [
  27         {
  28             u'url': u'http://www.hodiho.fr/2013/02/regis-plante-sa-jeep.html',
  29             u'file': u'13601338388002.mp4',
  30             u'md5': u'6e15c93721d7ec9e9ca3fdbf07982cfd',
  31             u'info_dict': {
  32                 u"uploader": u"www.hodiho.fr",
  33                 u"title": u"R\u00e9gis plante sa Jeep"
  34             }
  35         },
  36         # embedded vimeo video
  37         {
  38             u'add_ie': ['Vimeo'],
  39             u'url': u'http://skillsmatter.com/podcast/home/move-semanticsperfect-forwarding-and-rvalue-references',
  40             u'file': u'22444065.mp4',
  41             u'md5': u'2903896e23df39722c33f015af0666e2',
  42             u'info_dict': {
  43                 u'title': u'ACCU 2011: Move Semantics,Perfect Forwarding, and Rvalue references- Scott Meyers- 13/04/2011',
  44                 u"uploader_id": u"skillsmatter",
  45                 u"uploader": u"Skills Matter",
  46             }
  47         },
  48         # bandcamp page with custom domain
  49         {
  50             u'add_ie': ['Bandcamp'],
  51             u'url': u'http://bronyrock.com/track/the-pony-mash',
  52             u'file': u'3235767654.mp3',
  53             u'info_dict': {
  54                 u'title': u'The Pony Mash',
  55                 u'uploader': u'M_Pallante',
  56             },
  57             u'skip': u'There is a limit of 200 free downloads / month for the test song',
  58         },
  59         # embedded brightcove video
  60         # it also tests brightcove videos that need to set the 'Referer' in the
  61         # http requests
  62         {
  63             u'add_ie': ['Brightcove'],
  64             u'url': u'http://www.bfmtv.com/video/bfmbusiness/cours-bourse/cours-bourse-l-analyse-technique-154522/',
  65             u'info_dict': {
  66                 u'id': u'2765128793001',
  67                 u'ext': u'mp4',
  68                 u'title': u'Le cours de bourse : l’analyse technique',
  69                 u'description': u'md5:7e9ad046e968cb2d1114004aba466fd9',
  70                 u'uploader': u'BFM BUSINESS',
  71             },
  72             u'params': {
  73                 u'skip_download': True,
  74             },
  75         },
  76         # Direct link to a video
  77         {
  78             u'url': u'http://media.w3.org/2010/05/sintel/trailer.mp4',
  79             u'file': u'trailer.mp4',
  80             u'md5': u'67d406c2bcb6af27fa886f31aa934bbe',
  81             u'info_dict': {
  82                 u'id': u'trailer',
  83                 u'title': u'trailer',
  84                 u'upload_date': u'20100513',
  85             }
  86         }
  87     ]
  88
  89     def report_download_webpage(self, video_id):
  90         """Report webpage download."""
  91         if not self._downloader.params.get('test', False):
  92             self._downloader.report_warning(u'Falling back on generic information extractor.')
  93         super(GenericIE, self).report_download_webpage(video_id)
  94
  95     def report_following_redirect(self, new_url):
  96         """Report information extraction."""
  97         self._downloader.to_screen(u'[redirect] Following redirect to %s' % new_url)
  98
  99     def _send_head(self, url):
 100         """Check if it is a redirect, like url shorteners, in case return the new url."""
 101         class HeadRequest(compat_urllib_request.Request):
 102             def get_method(self):
 103                 return "HEAD"
 104
 105         class HEADRedirectHandler(compat_urllib_request.HTTPRedirectHandler):
 106             """
 107             Subclass the HTTPRedirectHandler to make it use our
 108             HeadRequest also on the redirected URL
 109             """
 110             def redirect_request(self, req, fp, code, msg, headers, newurl):
 111                 if code in (301, 302, 303, 307):
 112                     newurl = newurl.replace(' ', '%20')
 113                     newheaders = dict((k,v) for k,v in req.headers.items()
 114                                       if k.lower() not in ("content-length", "content-type"))
 115                     return HeadRequest(newurl,
 116                                        headers=newheaders,
 117                                        origin_req_host=req.get_origin_req_host(),
 118                                        unverifiable=True)
 119                 else:
 120                     raise compat_urllib_error.HTTPError(req.get_full_url(), code, msg, headers, fp)
 121
 122         class HTTPMethodFallback(compat_urllib_request.BaseHandler):
 123             """
 124             Fallback to GET if HEAD is not allowed (405 HTTP error)
 125             """
 126             def http_error_405(self, req, fp, code, msg, headers):
 127                 fp.read()
 128                 fp.close()
 129
 130                 newheaders = dict((k,v) for k,v in req.headers.items()
 131                                   if k.lower() not in ("content-length", "content-type"))
 132                 return self.parent.open(compat_urllib_request.Request(req.get_full_url(),
 133                                                  headers=newheaders,
 134                                                  origin_req_host=req.get_origin_req_host(),
 135                                                  unverifiable=True))
 136
 137         # Build our opener
 138         opener = compat_urllib_request.OpenerDirector()
 139         for handler in [compat_urllib_request.HTTPHandler, compat_urllib_request.HTTPDefaultErrorHandler,
 140                         HTTPMethodFallback, HEADRedirectHandler,
 141                         compat_urllib_request.HTTPErrorProcessor, compat_urllib_request.HTTPSHandler]:
 142             opener.add_handler(handler())
 143
 144         response = opener.open(HeadRequest(url))
 145         if response is None:
 146             raise ExtractorError(u'Invalid URL protocol')
 147         return response
 148
 149     def _real_extract(self, url):
 150         parsed_url = compat_urlparse.urlparse(url)
 151         if not parsed_url.scheme:
 152             self._downloader.report_warning('The url doesn\'t specify the protocol, trying with http')
 153             return self.url_result('http://' + url)
 154         video_id = os.path.splitext(url.split('/')[-1])[0]
 155
 156         try:
 157             response = self._send_head(url)
 158
 159             # Check for redirect
 160             new_url = response.geturl()
 161             if url != new_url:
 162                 self.report_following_redirect(new_url)
 163                 return self.url_result(new_url)
 164
 165             # Check for direct link to a video
 166             content_type = response.headers.get('Content-Type', '')
 167             m = re.match(r'^(?:audio|video)/(?P<format_id>.+)$', content_type)
 168             if m:
 169                 upload_date = response.headers.get('Last-Modified')
 170                 if upload_date:
 171                     upload_date = unified_strdate(upload_date)
 172                 assert (url_basename(url) == 'trailer.mp4')
 173                 return {
 174                     'id': video_id,
 175                     'title': os.path.splitext(url_basename(url))[0],
 176                     'formats': [{
 177                         'format_id': m.group('format_id'),
 178                         'url': url,
 179                     }],
 180                     'upload_date': upload_date,
 181                 }
 182
 183         except compat_urllib_error.HTTPError:
 184             # This may be a stupid server that doesn't like HEAD, our UA, or so
 185             pass
 186
 187         try:
 188             webpage = self._download_webpage(url, video_id)
 189         except ValueError:
 190             # since this is the last-resort InfoExtractor, if
 191             # this error is thrown, it'll be thrown here
 192             raise ExtractorError(u'Failed to download URL: %s' % url)
 193
 194         self.report_extraction(video_id)
 195
 196         # it's tempting to parse this further, but you would
 197         # have to take into account all the variations like
 198         #   Video Title - Site Name
 199         #   Site Name | Video Title
 200         #   Video Title - Tagline | Site Name
 201         # and so on and so forth; it's just not practical
 202         video_title = self._html_search_regex(
 203             r'(?s)<title>(.*?)</title>', webpage, u'video title',
 204             default=u'video')
 205
 206         # video uploader is domain name
 207         video_uploader = self._search_regex(
 208             r'^(?:https?://)?([^/]*)/.*', url, u'video uploader')
 209
 210         # Look for BrightCove:
 211         bc_url = BrightcoveIE._extract_brightcove_url(webpage)
 212         if bc_url is not None:
 213             self.to_screen(u'Brightcove video detected.')
 214             return self.url_result(bc_url, 'Brightcove')
 215
 216         # Look for embedded Vimeo player
 217         mobj = re.search(
 218             r'<iframe[^>]+?src="(https?://player.vimeo.com/video/.+?)"', webpage)
 219         if mobj:
 220             player_url = unescapeHTML(mobj.group(1))
 221             surl = smuggle_url(player_url, {'Referer': url})
 222             return self.url_result(surl, 'Vimeo')
 223
 224         # Look for embedded YouTube player
 225         matches = re.findall(
 226             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:www\.)?youtube\.com/embed/.+?)\1', webpage)
 227         if matches:
 228             urlrs = [self.url_result(unescapeHTML(tuppl[1]), 'Youtube')
 229                      for tuppl in matches]
 230             return self.playlist_result(
 231                 urlrs, playlist_id=video_id, playlist_title=video_title)
 232
 233         # Look for embedded Dailymotion player
 234         matches = re.findall(
 235             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:www\.)?dailymotion\.com/embed/video/.+?)\1', webpage)
 236         if matches:
 237             urlrs = [self.url_result(unescapeHTML(tuppl[1]), 'Dailymotion')
 238                      for tuppl in matches]
 239             return self.playlist_result(
 240                 urlrs, playlist_id=video_id, playlist_title=video_title)
 241
 242         # Look for embedded Wistia player
 243         match = re.search(
 244             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:fast\.)?wistia\.net/embed/iframe/.+?)\1', webpage)
 245         if match:
 246             return {
 247                 '_type': 'url_transparent',
 248                 'url': unescapeHTML(match.group('url')),
 249                 'ie_key': 'Wistia',
 250                 'uploader': video_uploader,
 251                 'title': video_title,
 252                 'id': video_id,
 253             }
 254
 255         # Look for embedded blip.tv player
 256         mobj = re.search(r'<meta\s[^>]*https?://api.blip.tv/\w+/redirect/\w+/(\d+)', webpage)
 257         if mobj:
 258             return self.url_result('http://blip.tv/seo/-'+mobj.group(1), 'BlipTV')
 259         mobj = re.search(r'<(?:iframe|embed|object)\s[^>]*https?://(?:\w+\.)?blip.tv/(?:play/|api\.swf#)([a-zA-Z0-9]+)', webpage)
 260         if mobj:
 261             player_url = 'http://blip.tv/play/%s.x?p=1' % mobj.group(1)
 262             player_page = self._download_webpage(player_url, mobj.group(1))
 263             blip_video_id = self._search_regex(r'data-episode-id="(\d+)', player_page, u'blip_video_id', fatal=False)
 264             if blip_video_id:
 265                 return self.url_result('http://blip.tv/seo/-'+blip_video_id, 'BlipTV')
 266
 267         # Look for Bandcamp pages with custom domain
 268         mobj = re.search(r'<meta property="og:url"[^>]*?content="(.*?bandcamp\.com.*?)"', webpage)
 269         if mobj is not None:
 270             burl = unescapeHTML(mobj.group(1))
 271             # Don't set the extractor because it can be a track url or an album
 272             return self.url_result(burl)
 273
 274         # Look for embedded Vevo player
 275         mobj = re.search(
 276             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:cache\.)?vevo\.com/.+?)\1', webpage)
 277         if mobj is not None:
 278             return self.url_result(mobj.group('url'))
 279
 280         # Start with something easy: JW Player in SWFObject
 281         mobj = re.search(r'flashvars: [\'"](?:.*&)?file=(http[^\'"&]*)', webpage)
 282         if mobj is None:
 283             # Broaden the search a little bit
 284             mobj = re.search(r'[^A-Za-z0-9]?(?:file|source)=(http[^\'"&]*)', webpage)
 285         if mobj is None:
 286             # Broaden the search a little bit: JWPlayer JS loader
 287             mobj = re.search(r'[^A-Za-z0-9]?file["\']?:\s*["\'](http[^\'"]*)', webpage)
 288         if mobj is None:
 289             # Try to find twitter cards info
 290             mobj = re.search(r'<meta (?:property|name)="twitter:player:stream" (?:content|value)="(.+?)"', webpage)
 291         if mobj is None:
 292             # We look for Open Graph info:
 293             # We have to match any number spaces between elements, some sites try to align them (eg.: statigr.am)
 294             m_video_type = re.search(r'<meta.*?property="og:video:type".*?content="video/(.*?)"', webpage)
 295             # We only look in og:video if the MIME type is a video, don't try if it's a Flash player:
 296             if m_video_type is not None:
 297                 mobj = re.search(r'<meta.*?property="og:video".*?content="(.*?)"', webpage)
 298         if mobj is None:
 299             # HTML5 video
 300             mobj = re.search(r'<video[^<]*(?:>.*?<source.*?)? src="([^"]+)"', webpage, flags=re.DOTALL)
 301         if mobj is None:
 302             raise ExtractorError(u'Unsupported URL: %s' % url)
 303
 304         # It's possible that one of the regexes
 305         # matched, but returned an empty group:
 306         if mobj.group(1) is None:
 307             raise ExtractorError(u'Did not find a valid video URL at %s' % url)
 308
 309         video_url = mobj.group(1)
 310         video_url = compat_urlparse.urljoin(url, video_url)
 311         video_id = compat_urllib_parse.unquote(os.path.basename(video_url))
 312
 313         # here's a fun little line of code for you:
 314         video_id = os.path.splitext(video_id)[0]
 315
 316         return {
 317             'id':       video_id,
 318             'url':      video_url,
 319             'uploader': video_uploader,
 320             'title':    video_title,
 321         }